Linux partition window

Introduction

The modern desktop PC has enabled computing to expand beyond the centralized mainframe model and has allowed both vendors and customers to concentrate their enterprises on low-priced, high-performance hardware platforms. However, the PC-based model is subject to less than ideal aspects of multivendor hardware/software/operating system architecture, particularly in the support of large, varied legacy and nonlegacy environments.1,2 Organizations attempt to overcome these potential problems by selecting hardware, software, and operating systems they can support on a daily basis. This uniformity constraint is a natural consequence of the negative aspects of the heterogeneous computing environment. However, there are many existing applications that fall outside this constraint and need individual support, such as those found in special purpose laboratories.3 Support of these applications forces the laboratory into unique configurations not easily achieved through standard information technology (IT) practices and system administration.3 Consequently, support of special purpose engineering technology laboratories often becomes the responsibility of laboratory personnel.

In the laboratory environment, the resources available for system support may involve a mystifying number of unique software applications from a variety of vendors. In some cases these programs perform their tasks well; in other cases they are not always ideally matched to the needed tasks. Laboratory support tasks may include system generation or repair, software installation or repair, and system backup. The topics of system and software repair are critical issues in the field of high-availability (HA), fault-tolerant (FT), and mission-critical computing. There are existing architecture schemes that limit exposure to many types of system failures, thus easing system administration and maintenance. However, such schemes require a significant investment and are most likely found in mid- to large-scale telecommunication, networking, and server applications rather than in the laboratory. Further, typical HA or FT architecture constraints may force exclusion of vital and unique laboratory hardware or software or require their relocation to remote server systems.

For modest non-HA, non-FT-based machines, system support including file recovery and physical recovery can be accomplished with the current generation of Linux distributions coexisting with Windows in a dual-boot environment.4,5 These Linux distributions use a widely implemented open operating system, available either free or for a nominal charge, that competes with Windows OS products, especially Windows NT.be Linux provides a secure system environment that features password-protected user logins, individually assigned user capability by both name and group declarations, and directories and files with accessibility keyed to these parameters. Most importantly, for system maintenance, Linux is designed for complete compatibility with the Windows filesystem architecture. The filesystem compatibility also extends to Windows NT, Macintosh, OS/2, and network configurations.

Engineering technology professionals competent in Unix or Windows NT should feel comfortable navigating a new Linux installation. Users whose experience is limited to a successful Windows 9x background will face a modest learning curve, but extensive Linux literature, both in print and on the Web, can significantly expedite the learning process. To other users, the time invested in learning a new operating system must be weighed against potential gains in PC laboratory administration and maintenance. It is noted that if current IT laboratory support is satisfactory, the potential productivity gain from Linux maintenance operations will be slight. For this case the potential benefit of the dual-boot environment lies in the use of Linux as a direct laboratory tool.

Sharing Disk Drive Resources: Microsoft Windows 9x and Linux

Linux can be easily installed onto platforms preloaded with Microsoft Windows 9x. There are two primary mechanisms to accomplish this task: installation of Linux to an unused partition, or installation of Linux within a native Windows filesystem.8,9.10 If these mechanisms are impractical, Linux may be installed on a supplemental hard drive.

PCs allow a maximum of four partitions for each physical hard drive. One can be designated as a primary DOS (Windows) partition, and another may be designated an extended DOS partition, commonly written EXT DOS. The extended partition is constructed to support multiple subpartitions, known as logical drives.11 Windows restricts itself to these two partition types, with a maximum of one DOS and one EXT DOS partition per hard drive. Linux requires a minimum of two partitions: one for the operating system's file space and one for the operating system's swap space. The swap space partition is small and is reserved for caching memory contents to disk. Its size is computed as a multiple of the installed system memory, with the value of the multiplier generally 2 or 3.

Linux also allows the use of additional partitions on a hard drive. Like Windows, Linux distributions support four partitions, referred to as primary partitions, and can support additional partitions through the use of extended partitions. Linux distributions can recognize and interact with DOS and EXT DOS partitions, while Windows can recognize the existence of non-DOS partitions but cannot access them without third-party software.

Site DIsk Space ASUSSMet

To partition disks with Linux, the system hard drive or drives must have unused space that exists in one or more of the following forms: (1) existing drive space never allocated to DOS or EXT DOS partitions, (2) free space within an existing DOS or EXT DOS partition, or (3) space within new hard drives.12 The amount of unused space needed by Linux depends on the distribution, but generally falls in the range of 300 MB to 1300 MB. To use Linux for system maintenance, free space ranging from 300 MB to 600 MB is satisfactory. Beyond the 600 MB boundary, Linux distributions add many native operating system features including programming, database and office management, web authoring, networking, and graphics applications and development tools.

Mature Windows-based computers often contain unallocated disk space, because as maximum disk space provided by original equipment is approached, newer drives with dramatically larger capacity are substituted. Manufacturers now produce hard drives capable of meeting storage demands in the giga- and terrabyte range. For typical small laboratory applications, the capacities of these drives significantly exceed actual storage requirements of the Windows-based computers.

The presence of unused space on these machines can manifest itself in two ways: (1) a partition table that allocates 100% of disk space to DOS and EXT DOS partitions, resulting in a Windows filesystem that contains a huge amount of free space, or (2) a partition table that fails to allocate 100% of disk space to DOS and EXT DOS partitions. Often the second manifestation is due to BIOS, or native operating system limitations encountered when the hard drive was partitioned.13)14 Depending on the source, these limits may occur at 2-, 4- and 8-GB boundaries.15 In both cases the unused space is available to Linux.

If all drive space has been allocated into DOS and EXT DOS partitions and the filesystem contains enough free space for a Linux installation, the partition layout (as denoted in the system partition table) may be reworked for dual-boot operation. There are common Window-based utilities supplied with Linux distributions and commercial third-party Windows applications that are capable of performing the repartitioning task. These utilities can split and relocate DOS and EXT DOS partitions, thereby creating the empty space necessary for Linux installation. If repartitioning the existing drive or adding another drive are not options, certain Linux distributions allow for Linux installation directly into the Windows filesystem.16,17 The filesystem environment becomes a pseudo Linux filesystem constructed from a Windows-based Linux image file instead of a partition-based native Linux system. Successful Linux operation in this configuration relies on ample amounts of memory to speed system operation.